Updated August 23, 2026 · 13 min read · by Sam Smith
Kalshi lists the same fifty college football programs on two different boards this season. The first asks whether a team makes the twelve-team College Football Playoff. The second asks whether it reaches the semifinal, the round of four that plays on January 14 and 15, 2027.
Twelve slots on one board, four on the other. Same fifty names, same season, same traders. (New to event contracts? Our guide to college football on Kalshi covers how these markets settle.)
Add up every price on the qualifying board and you get 1,231.5 cents against a true total of 1,200. That is about as tight as a fifty-outcome market gets. Add up every price on the semifinal board and you get 585 cents against a true total of 400.
A 46% overshoot on one board and a 2.6% overshoot on the other, from the same crowd, in the same week. The reason turns out to have almost nothing to do with football, and sorting it out is what makes the rest of the board readable.
Two Boards, The Same Fifty Teams
Reaching a semifinal requires making the Playoff first. A team that misses the field cannot win a quarterfinal. So every program's semifinal probability is capped by its qualifying probability, on every board, always. When a market breaks that rule, something in the pricing is wrong.
The mechanism separating the two boards is seeding. Under the 2026-27 format, confirmed by the College Football Playoff's own announcement, the field stays at twelve and the four highest-ranked teams take seeds one through four along with first-round byes. Conference championship status affects access, not seed, a change from the 2024 season when the byes were reserved for the top four conference champions.
That produces two very different paths. A team seeded in the top four sits out the first round and needs to win a single game, its quarterfinal, to reach a semifinal. A team seeded fifth through twelfth has to win two: a first-round game, then a quarterfinal. Seeds five through eight host that opener on campus; seeds nine through twelve travel.
So the semifinal board asks a different question from the qualifying board: who lands a bye. Two teams equally likely to make the field can sit a long way apart here.
The contract language is plain. Every market on the board reads, with only the team name changing: "If USC qualifies for the College Football Playoff Semifinals in the 2026-27 season, then the market resolves to Yes." The four semifinalists are settled by the quarterfinals on December 30 and January 1, so the outcome is known on New Year's Day even though Kalshi lists an expected expiration of January 24, 2027.
Free: The Weekly PM Market Brief — the 8-model panel's graded record, the week's biggest market-vs-model gaps, and what's spiking next. One email, Sundays.
The 159 Cents Nobody Has Ever Traded
Twenty-eight of the fifty semifinal contracts have never traded a single time. Not once, since the board opened on June 24.
All fifty carry a bid and an ask, so a screen shows a price for every team. But those twenty-eight sit at the one- or two-cent floor against asks running from nine to fifteen cents, and they hold zero open interest between them. Each one therefore shows a midpoint somewhere between five and eight and a half cents. Added up, those twenty-eight midpoints come to 159 cents of the board's total, conjured entirely out of the gap between a floor bid and a speculative offer.
Strip them out. The twenty-two contracts that have actually traded sum to 426 cents against a true total of 400, an overshoot of 6.5%. The qualifying board, where all fifty contracts have traded, runs at 2.6% with a median spread of four cents. The semifinal board's median spread is nine cents.
The 46% figure measures thinness. On a contract with no trades behind it, a midpoint carries no information about anything; it is the middle of an empty room.
That thinness produces four rows that cannot be true. Arkansas is quoted higher to reach the semifinal than to make the Playoff at all, a 240% conditional. North Texas prices at 143%, Tulane at 125%, Kentucky at 122%. Each is an arithmetic impossibility, and each exists because a floor-level bid and a wide ask average out to something larger than a real quoted price on the other board.
- Arkansas — semifinal 1/11¢ against qualify 1/4¢, an implied 240%. Selling the semifinal at the bid and buying the qualifier at the ask costs 3¢.
- North Texas — semifinal 1/9¢ against qualify 1/6¢, an implied 143%. Selling the semifinal at the bid and buying the qualifier at the ask costs 5¢.
- Tulane — semifinal 1/9¢ against qualify 2/6¢, an implied 125%. Selling the semifinal at the bid and buying the qualifier at the ask costs 5¢.
- Kentucky — semifinal 1/10¢ against qualify 2/7¢, an implied 122%. Selling the semifinal at the bid and buying the qualifier at the ask costs 6¢.
A semifinalist is by definition also a qualifier, so no row here can price above 100%. Prices as of August 23, 2026.
Every one of the four costs money to trade. That is precisely why nobody has corrected them: the contradiction is visible on a screen and unreachable in a book.
One more depth note, because the headline figure flatters this board badly. The semifinal market shows 487,539 contracts of lifetime volume against 486,266 of open interest. A single contract, LSU, accounts for 96.5% of it. The other forty-nine teams have traded roughly 16,900 contracts between them since June.
Ohio State, Oregon And The 1-For-8 Bye
A first-round bye is supposed to be the prize. Six AI models priced this board for us, and all six built their numbers the same way: the chance of a top-four seed multiplied by a one-game conversion rate near 57 to 60 percent, plus smaller contributions from the two-win paths. The format is designed to reward exactly that: the bye exists to make a top-four seed's road shorter by a full game.
The two completed brackets say something else.
In the 2024 season, the four bye teams were Oregon, Georgia, Boise State and Arizona State. All four lost their quarterfinals. Penn State beat Boise State 31-14, Texas beat Arizona State 39-31, Ohio State beat Oregon 41-21, and Notre Dame beat Georgia 23-10. All four semifinalists came from the group that had played a first-round game a week earlier.
A year later the byes went to Georgia, Indiana, Ohio State and Texas Tech. Only Indiana converted. Miami beat Ohio State 24-14, Ole Miss beat Georgia 39-34, and Oregon beat Texas Tech 23-0.
Across both seasons, one of eight bye teams has reached a semifinal. Every quarterfinal pairs a rested top seed against a first-round winner, so across those eight games the bye side went 1-7.
The bye has been worth one semifinal in eight tries. Both completed brackets, eight rested top-four seeds, one survivor. Every quarterfinal is a rested team against a team that just won a playoff game, and the rested team has lost seven of them.
Two of the six models flagged the tension themselves. Claude Sonnet held its top tier deliberately flat because of it, arguing the bracket log "is a real (if small-sample) signal that quarterfinal upsets are common." Claude Fable put it more bluntly, rejecting the higher ceilings on the grounds that "a bye is worth one game, not a coronation, so no preseason team deserves more than ~38-40."
Eight games is not a law, and the bye teams are by construction the four teams a committee ranked highest, which should make them win more often, not less. Both readings stay open. But any model pricing a bye as a near-automatic semifinal berth is pricing against the only on-field evidence this format has produced.
Where The Panel And The Market Actually Disagree
None of those six models saw a market price at any point. Each one got the same deterministically fetched card: the verified 2026-27 seeding and bye rules, both completed twelve-team brackets game by game, every listed program's full 2025 record, its conference and head coach, and both 2026 preseason polls. Each model had to distribute probability across all fifty names in a single pass, summing to 400 rather than 100, because exactly four teams get there. Then every model read the others' reasoning anonymously and could revise. The second-round numbers below are the ones we publish. Every model's own board was renormalized to 400 before blending; the panel figure on each row is the median of those six, which is why the published column lands at 400.5 rather than dead on 400. We publish the median of the six rather than promoting any single model's board.
Six of the panel's eight seats completed this board. The other two returned boards that did not respect the four-slot constraint, one summing far above 400 and one far below, and were rejected rather than rescaled. Their numbers appear nowhere on this page.
Two market columns follow, and they are not the same number. The fifty-team board below shows Kalshi's raw midpoint, phantoms included, because that is what a screen actually displays. The seat table's cleaned column takes the twenty-two traded contracts and renormalizes them from 426 to 400, which is the only like-for-like comparison available. Judging the panel against the raw 585 would manufacture a fade on almost every team on the page.
The Board
| Team | Conference | Kalshi mid | Bid / ask | Ever traded | Panel |
|---|---|---|---|---|---|
| Ohio St. | Big Ten | 38.5¢ | 35 / 42 | yes | 36.0% |
| Oregon | Big Ten | 39.0¢ | 37 / 41 | yes | 33.4% |
| Georgia | SEC | 33.5¢ | 29 / 38 | yes | 32.2% |
| Notre Dame | FBS Indep. | 38.0¢ | 34 / 42 | yes | 29.6% |
| Texas | SEC | 32.0¢ | 26 / 38 | yes | 27.9% |
| Indiana | Big Ten | 33.0¢ | 31 / 35 | yes | 25.9% |
| Miami (FL) | ACC | 37.0¢ | 31 / 43 | yes | 23.3% |
| Texas A&M | SEC | 13.5¢ | 9 / 18 | yes | 17.1% |
| Ole Miss | SEC | 13.5¢ | 10 / 17 | yes | 15.1% |
| Oklahoma | SEC | 19.5¢ | 18 / 21 | yes | 13.8% |
| Alabama | SEC | 13.5¢ | 8 / 19 | yes | 12.7% |
| LSU | SEC | 17.5¢ | 15 / 20 | yes | 11.3% |
| Texas Tech | Big 12 | 17.0¢ | 13 / 21 | yes | 10.3% |
| USC | Big Ten | 11.5¢ | 9 / 14 | yes | 9.1% |
| BYU | Big 12 | 10.0¢ | 5 / 15 | yes | 8.0% |
| Michigan | Big Ten | 10.5¢ | 4 / 17 | yes | 8.0% |
| Penn St. | Big Ten | 15.5¢ | 14 / 17 | yes | 6.3% |
| Tennessee | SEC | 8.5¢ | 2 / 15 | never | 6.0% |
| Washington | Big Ten | 8.0¢ | 2 / 14 | yes | 6.0% |
| Utah | Big 12 | 6.5¢ | 1 / 12 | never | 5.8% |
| SMU | ACC | 7.5¢ | 2 / 13 | never | 5.0% |
| Clemson | ACC | 6.5¢ | 1 / 12 | never | 4.3% |
| Louisville | ACC | 7.5¢ | 2 / 13 | never | 4.1% |
| Iowa | Big Ten | 6.0¢ | 1 / 11 | never | 4.0% |
| Houston | Big 12 | 5.5¢ | 1 / 10 | never | 4.0% |
| Missouri | SEC | 7.5¢ | 2 / 13 | never | 3.8% |
| Illinois | Big Ten | 5.0¢ | 1 / 9 | never | 3.3% |
| Vanderbilt | SEC | 5.5¢ | 1 / 10 | never | 3.0% |
| TCU | Big 12 | 5.5¢ | 1 / 10 | never | 2.9% |
| Arizona | Big 12 | 5.0¢ | 1 / 9 | never | 2.5% |
| Virginia | ACC | 5.5¢ | 1 / 10 | never | 2.5% |
| Georgia Tech | ACC | 5.0¢ | 1 / 9 | never | 2.3% |
| North Texas | American | 5.0¢ | 1 / 9 | never | 2.2% |
| Florida | SEC | 8.5¢ | 3 / 14 | yes | 2.0% |
| South Carolina | SEC | 4.0¢ | 2 / 6 | yes | 1.9% |
| Tulane | American | 5.0¢ | 1 / 9 | never | 1.9% |
| James Madison | Sun Belt - East | 5.0¢ | 1 / 9 | never | 1.7% |
| Pittsburgh | ACC | 5.5¢ | 1 / 10 | never | 1.5% |
| North Carolina St. | ACC | 5.0¢ | 1 / 9 | never | 1.2% |
| Arizona St. | Big 12 | 5.0¢ | 1 / 9 | never | 1.2% |
| Auburn | SEC | 6.5¢ | 1 / 12 | yes | 1.1% |
| Florida St. | ACC | 5.0¢ | 1 / 9 | never | 1.1% |
| Kansas St. | Big 12 | 5.5¢ | 1 / 10 | never | 0.9% |
| Baylor | Big 12 | 5.0¢ | 1 / 9 | never | 0.9% |
| Virginia Tech | ACC | 5.5¢ | 1 / 10 | never | 0.8% |
| California | ACC | 5.0¢ | 1 / 9 | never | 0.8% |
| Kentucky | SEC | 5.5¢ | 1 / 10 | never | 0.6% |
| Oklahoma St. | Big 12 | 5.0¢ | 1 / 9 | never | 0.4% |
| Miami (OH) | MAC | 5.0¢ | 1 / 9 | never | 0.4% |
| Arkansas | SEC | 6.0¢ | 1 / 11 | yes | 0.4% |
All fifty contracts carry both a bid and an ask; the twenty-eight that have never traded sit at a one- or two-cent floor. Kalshi mid is the raw bid-ask midpoint, phantoms included, shown to one decimal so the renormalization in the next table can be reproduced from this column. The panel column is the median of six model estimates, each renormalized to the four-slot total of 400. Prices as of August 23, 2026.
Every Seat's Number
| Team | Claude Fable | Claude Opus | Claude Sonnet | GPT | Gemini | Kimi | Panel | Market (cleaned) | Gap |
|---|---|---|---|---|---|---|---|---|---|
| Ohio St. | 38.3 | 35.0 | 34.5 | 35.5 | 36.5 | 38.9 | 36.0 | 36.2 | -0.2 |
| Oregon | 34.3 | 32.5 | 29.5 | 32.0 | 34.5 | 34.4 | 33.4 | 36.6 | -3.2 |
| Georgia | 33.3 | 31.0 | 28.4 | 31.0 | 33.5 | 33.4 | 32.2 | 31.5 | +0.7 |
| Notre Dame | 30.2 | 28.5 | 27.4 | 29.0 | 31.5 | 31.9 | 29.6 | 35.7 | -6.1 |
| Texas | 29.2 | 26.5 | 25.4 | 26.0 | 30.5 | 29.4 | 27.9 | 30.0 | -2.1 |
| Indiana | 26.2 | 25.5 | 22.4 | 23.5 | 26.5 | 26.4 | 25.9 | 31.0 | -5.1 |
| Miami (FL) | 24.2 | 22.5 | 20.3 | 22.0 | 24.5 | 24.9 | 23.3 | 34.7 | -11.4 |
| Texas A&M | 17.1 | 16.5 | 16.3 | 17.0 | 17.5 | 18.0 | 17.1 | 12.7 | +4.4 |
| Ole Miss | 15.1 | 14.0 | 14.2 | 15.0 | 16.5 | 16.0 | 15.1 | 12.7 | +2.4 |
| Oklahoma | 13.6 | 13.0 | 12.2 | 14.0 | 15.5 | 14.0 | 13.8 | 18.3 | -4.5 |
| Alabama | 12.1 | 12.5 | 9.1 | 13.0 | 13.5 | 13.0 | 12.7 | 12.7 | +0.0 |
| LSU | 12.1 | 11.5 | 11.2 | 11.0 | 12.5 | 11.0 | 11.3 | 16.4 | -5.1 |
| Texas Tech | 8.1 | 11.0 | 10.2 | 10.0 | 10.5 | 12.0 | 10.3 | 16.0 | -5.7 |
| USC | 9.1 | 9.5 | 9.1 | 9.0 | 9.5 | 9.0 | 9.1 | 10.8 | -1.7 |
Every seat on this panel is graded against real market settlements — records to date: Claude Fable 82% on 616 graded calls · Claude Opus 83% on 686 graded calls · Claude Sonnet 81% on 664 graded calls · GPT 84% on 5,836 graded calls · Gemini 86% on 4,570 graded calls · Kimi 82% on 2,442 graded calls. Recomputed daily; the full scoreboard is public.
Model estimates generated August 23, 2026, price-blind. These are model estimates, not predictions of fact and not financial advice or trading advice. Models are frequently wrong, and the market price reflects real traders' money. Kalshi is a CFTC-regulated exchange; 18+, availability varies by state. Every number here gets graded in public once this market settles, on the full graded scoreboard.
More live boards from the same panel: College Football Playoff Odds 2026, the twelve-team qualifying board this page is built against, where Notre Dame leads at 81-84¢ · College Football Championship Odds, the title board, where Ohio State, Notre Dame and Texas share the lead near 12¢ · College Football Undefeated Odds, where Texas Tech leads the field at 34¢. Prices fetched August 23, 2026.
Once the phantom contracts come out, the top of the board is close to a handshake. The panel puts Ohio State at 36.0 against a cleaned market price of 36.2. Georgia lands at 32.2 against 31.5. Texas sits 2.1 points under the crowd and Oregon 3.2 under, close enough that neither is the story.
The disagreements are concentrated in six names, and Miami is the largest of them. Kalshi's semifinal market quotes the Hurricanes 31 bid, 43 ask, which cleans to 34.7 and makes them the board's fourth-likeliest semifinalist. Six seats put them seventh, at 23.3, a gap of 11.4 points, and every one of the six models landed between 20.3 and 24.9. Miami reached the final last season, losing 21-27 to Indiana, and the market appears to be paying for that run more than the models will.
Penn State is the next largest gap at 8.3 points, though it carries a caveat the Miami row does not: only 167 semifinal contracts have ever traded on Penn State, against 38,697 on its qualifying market. Notre Dame comes in 6.1 points light, Texas Tech 5.7, Indiana and LSU 5.1 each.
Only three teams on the whole board come out above the market, and only two by a margin worth the name. Texas A&M is the clearest at plus 4.4, with all six models between 16.3 and 18.0 against a cleaned market price of 12.7. Ole Miss gets plus 2.4. Georgia edges it by 0.7, and Alabama lands exactly level.
Dark Horses The Panel Won't Dismiss
Below ten percent the board thins out fast, but three names drew real numbers rather than rounding errors. USC sits at 9.1 with the tightest agreement on the entire board, six models between 9.0 and 9.5. BYU lands at 8.0 and Michigan at 8.0, both a couple of points under the market. Utah, Iowa and Missouri each had at least one model price them above 6, and Missouri drew the widest single-seat disagreement of the group, from 2.5 up to 7.6. The same panel scores the individual-award market on Heisman Trophy odds, where several of these programs carry a candidate.
Texas Tech Is The Board's Strangest Number
On the qualifying board, Texas Tech is quoted 68-69 cents, sixth of fifty and a rung above Georgia. On the semifinal board it is 13 bid, 21 ask. Divide one by the other and the market gives the Red Raiders roughly a 25% chance of surviving to the semifinal given they qualify.
Every other top-eight qualifier on the board sits between 46% and 53%. Texas Tech is priced at half the rate of its peers.
The explanation is structural, and adding the board up by conference shows it plainly.
| Conference | Teams listed | Sum of qualify prices | Sum of semifinal prices | Implied conversion |
|---|---|---|---|---|
| SEC | 15 | 410.0¢ | 195.0¢ | 48% |
| Big Ten | 9 | 342.5¢ | 167.0¢ | 49% |
| ACC | 11 | 191.0¢ | 95.0¢ | 50% |
| Big 12 | 10 | 184.0¢ | 70.0¢ | 38% |
Conferences with at least two listed teams and meaningful quoted size. The conversion column divides a conference's summed semifinal prices by its summed qualifying prices. Prices as of August 23, 2026.
The Big 12 converts a Playoff berth into a semifinal at 38% in the market's own pricing, roughly ten points below every other power conference. A Big 12 champion at 12-1 gets into the field on its conference's automatic bid, then ranks somewhere around fifth to eighth, which means no bye and two games to win instead of one.
Kimi, pricing blind, named the same team without being shown any of it: "The program the public most misprices on THIS board is Texas Tech: a near-lock to qualify as Big 12 favorite, but a 12-2 Big 12 champion ranks ~5-8, not top four, so it needs two playoff wins; my 12 is roughly a third of its qualifier price."
There is a recent precedent, and it is Texas Tech's own. The Red Raiders won the Big 12 to reach 12-1 last season, earned a first-round bye with it, then lost their quarterfinal to Oregon 0-23 in the Orange Bowl to finish 12-2. In the twelve-team era, no Big 12 team has reached a semifinal.
One live variable sits outside all of this. Texas Tech quarterback Brendan Sorsby was ruled ineligible by the NCAA over sports wagering, and on June 8, 2026 a Lubbock County judge granted him a temporary injunction, ESPN reported, finding he would suffer "probable, imminent and irreparable injury" if kept off the field. He is eligible under that order but must sit the season's first two games. The NCAA filed an accelerated appeal the same day, and the underlying case is not scheduled to be tried until after the season. The panel never saw any of it; the data card carries no roster or eligibility facts by design.
Where The Panel Changed Its Mind
The revision round moved this board more than most. Agreement on Ohio State tightened from a twenty-point spread across six models to 4.4 points; Oregon went from nineteen points to five.
Gemini made the largest move on the board, cutting Ohio State from 52.8 to 36.5, Georgia from 45.9 to 33.5 and Oregon from 46.8 to 34.5. It adopted the seed-decomposition method the other seats were using: "I agree with Forecaster E's methodology of decomposing qualification probability into top-4, 5-8, and 9-12 buckets, because a top-4 seed requires one win, while 5-12 requires two."
Kimi took a correction it had not priced in round one, shading five ranked programs for coaching turnover: "Penn St. (Campbell), Michigan (Whittingham), LSU (Kiffin), Utah (Scalley), and Ole Miss (Golding) are all first-year staffs at ranked programs, so I priced each below raw poll position — Penn St. hardest at 6 despite playoff-caliber talent."
Claude Opus moved up rather than down, raising Oregon 4.7 points, Georgia 4.2 and Notre Dame 3.7, and credited a specific argument for the last one: Notre Dame "faces far less seed competition" and, with no conference championship game, avoids a thirteenth result that often drags an 11-1 team out of top-four consideration.
One seat barely moved. Claude Fable held Texas Tech near 8 through the revision round and held its ceiling on the favorites, and it is the model furthest below the market on the Red Raiders as a result.
What Would Change The Panel's Mind
The season has not started. Zero FBS games had been played as of this writing, and the first kick off on August 29, which means every number on this page is a projection built on last season's results and two preseason polls.
Four things would move it, each checkable against a date.
The first Saturday, August 29. The board's top seven are separated by about twelve points. A single loss by Ohio State, Oregon or Notre Dame in September resets the bye race, and the bye race is the board.
Conference championship weekend, early December. The last resume results before the committee sets its final rankings land here. Under straight seeding a title does not buy a bye, which is exactly the trap for the ACC and Big 12 champions: strong enough to qualify, rarely ranked high enough to rest. A Big 12 title for Texas Tech confirms the market's 25% conditional rather than refuting it.
Selection Day, December 6. The bracket, all twelve seeds and the four first-round matchups are announced together. Every conditional on this page collapses into a known path that morning.
The quarterfinals, December 30 and January 1. The board settles on New Year's Day. A fourth straight quarterfinal sweep by the first-round winners would take the bye side to 1-11 and make the panel's seed-conversion assumption untenable.
The Sorsby appeal is a fifth trigger with no fixed date attached, and it points at one team rather than the board.
Settlement timeline
| Date | What happens |
|---|---|
| August 29, 2026 | Season opens; first real information |
| December 6, 2026 | Selection Day: all twelve seeds and the four byes named |
| December 18-19, 2026 | First round, campus sites |
| December 30, 2026 And January 1, 2027 | Quarterfinals; the four semifinalists are known |
| January 14-15, 2027 | Semifinals: Orange Bowl, then Sugar Bowl |
| January 25, 2027 | Championship game, Allegiant Stadium, Las Vegas |
Kalshi lists an expected expiration of January 24, 2027, more than three weeks after the outcome is decided. The board re-scores on this page whenever the story moves, with the as-of stamp above.
The panel's calls are on the record. Our AI model panel puts a graded, public verdict on the twelve-team qualifying board these fifty teams are priced against: see the live model verdict on College Football Playoff Odds 2026.
Hottest Prediction Markets Right Now
- 2028 Democratic presidential nominee · $190M traded
- 2027 Pro Football Champion · $70M traded
- 2028 U.S. Presidential Election winner? · $62M traded
- 2028 Republican presidential nominee · $59M traded
- Pro Baseball Champion · $59M traded
- Bitcoin price at the end of 2026 · $32M traded
Every market above links to our full AI model verdict; browse them all on the OddsShopper prediction markets hub, and see how every settled call actually scored on the full graded scoreboard.
- How To Bet College Football On Kalshi: Markets & Value
- \ud83c\udfc8 College Football Week 0 Odds: All 8 Games, Priced Early
- College Football Betting Tips: 3 Edges The Market Is Slow To Price
- Kalshi College Football Picks This Week: Every Board, Scored
- College Football Parlays: Why Most Lose, Build Smarter
The Bottom Line
The semifinal board looks broken and mostly is not. Twenty-eight contracts nobody has traded account for 159 of the 185 cents it carries above a true total of 400, and the four rows that read as arithmetic impossibilities all cost money to trade. Cleaned up, the market runs 6.5% rich, which is a thin book behaving like a thin book.
What survives the cleanup is worth more than the headline. Panel and crowd agree almost exactly on Ohio State and Georgia, and split hard on six, with Miami at 23.3 against a cleaned 34.7 as the widest. Texas Tech gets a quarter of its qualifying price on a board where its peers get half, for a reason the conference table makes visible.
And whichever four teams get to Miami Gardens and New Orleans in January, the bye they earned in December has been worth one semifinal in eight tries. That record is two seasons old and may not survive a third. It is also the only record this format has.
